Documentation - Skills
GoodLads is not one model with one prompt. It is a set of skills, one per lever, each reading a specific slice of your Google Ads account and writing back a specific kind of change. A skill proposes something only when the evidence supports it - returning nothing is a normal, common result.
"Is this campaign capped, and would a lower cost target let the same budget buy more conversions?"
ReadsDay-by-day spend, the budget-vs-rank diagnostic, where spend is distributed and lost, ad-group performance, and what similar campaigns already learned.
WritesBidding-target test - a target-CPA or target-ROAS experiment run on a copy of the campaign.
"Is this campaign limited by its budget right now?"
Connect your account"Within this campaign, which ad groups deserve a different bidding target?"
ReadsAd-group performance, day-by-day spend, the budget diagnostic, and keyword quality.
WritesAd-group bid-target test - the same bidding-target experiment, scoped to one ad group.
"Is this App campaign's live bidding target still the right one?"
ReadsDay-by-day spend, how spend is distributed, and what similar campaigns already learned.
WritesApp campaign bid target - a live edit to the bidding target. App campaigns can't run our experiments at all, so this is the only path.
"Are some locations converting far better than others, and should the bid say so?"
ReadsPerformance by location, the budget diagnostic, day-by-day spend, and ad-group performance.
WritesLocation bid test - a bid adjustment experiment for a specific location.
"Do the account's best keywords have an ad written for them?"
ReadsDeep company research, keyword performance and quality, the ad copy running today, search terms, and which messaging angles are still available.
WritesNew ad - adds a new ad alongside your current ones, created paused.
"Which of my best keywords don't have an ad written for them yet?"
Connect your account"Is there a messaging angle this account hasn't tried yet that's worth a real test?"
ReadsDeep company research, the ad copy running today, and which angles this account has and hasn't tried.
WritesMessaging-angle ad test - new ad copy tested as an experiment against what's running today.
"Is there a messaging angle we haven't tested yet?"
Connect your account"Is one ad group serving several different search intents that deserve their own ad?"
ReadsKeyword-level intent detection inside each ad group, keyword quality, and the ad copy running today.
WritesAd-group split test - splits a mixed-intent ad group into focused ones and tests the split.
"Is any ad group in this campaign mixing different search intents?"
Connect your account"Is there proven demand this account can't serve well from the structure it has today?"
ReadsDeep company research, the landing page, search terms and keyword quality across the account, and what's already been tried.
WritesNew Search campaign - builds a whole new Search campaign, created paused.
"Is there search demand we're not serving well with a dedicated campaign?"
Connect your account"Which search terms are taking real money and returning nothing?"
ReadsSearch terms and where they route, account-wide.
WritesNegative keywords - blocks wasted search terms on the live campaign or ad group. No experiment form exists for this one.
"Which search terms are wasting spend in this account?"
Connect your account"Should an app you already advertise get its own install campaign?"
ReadsThe apps you already advertise elsewhere in the account.
WritesNew App campaign - builds a new App install campaign, created paused, for an app you already advertise.
"Can we launch an app install campaign for an app we already advertise?"
Connect your account"Is this capped campaign's budget worth raising, or would the money be better spent on efficiency first?"
ReadsThe budget-vs-rank diagnostic and spend distribution we already read, plus a sizing estimate of what the extra budget would be worth, still to come.
WritesNot built yet - would test a higher daily budget against the current one, split-tested rather than a raw raise.
Not built yet.See how these skills sit inside the wider account journey on Jobs to be done, the full read catalogue on Insights, or every write action on its own on Actions.